论文部分内容阅读
【Abstract】With the further deepening of globalization， frequent international exchanges and cooperations between countries， the workload of translators has become extremely large， and translation works are required to be quick and accurate. Human translators are overwhelmed by the new concepts that are emerging and the repetitive translation of academic and professional terms made the translators feel tired and exhausting. In recent years， the trials that machine translation assisted manual translation make people think that the quality of machine translation have been greatly improved， and people’s demands for machine translation have increased significantly. Therefore， this article will take Google translation， which is the most ideal translation tool till now and Youdao translation， which is loved by the young people as examples and explore China’s future development of machine translation and give advice for the future researches.
【Key words】Machine Translation； Google Translation； Youdao Translation
1. Introduction
As human beings become more and more comprehensive about information， the quality of the human translation has also been greatly improved. However， with the explosive growth of the amount of information， the number of texts that need to be translated has also multiplied， and the supply of human translators is far from being able to meet the translation needs of massive translated texts. With the continuous efforts of the machine translation researchers over the years， the quality of machine translation is roughly the same as that of human beings， so the people’s confidence in using machine translation has been greatly improved. However， due to the fact that the current machine translation can not reasonably adjust the order of words and divide the sentence components， there is a big gap between machine translation and users’ expectations. Thus， the research on the possibility of machine translation is conducive to the development of machine translation to break through the bottleneck and achieve better development situation. It also has important a practical significance for the further exploration of translation field.
2. The Feasibility of Google Translation and Youdao Translation
The development of machine translation so far has made great progress， but there are also many deficiencies. Below will use Google translation and Youdao translation as examples to discuss the feasibility of the current machine translation. 2.1 The Advantages and Disadvantages of Google Translation
As the benchmark for the machine translation industry， Google Translation initially adopted the most popular mainstream method， Phrased-Based Machine Translation （PBMT）， which firstly cuts the sentence into phrases， and then matches the phrases from the prepared statistical database， find the most similar text， and then these phrases will be rearranged their orders in accordance with the target language’s syntax， then finally get the translation. The translations translated in this way often have the following problems：
Firstly， the translations are rigid， and even have syntactic confusion. Its translation method is to split the sentence into several fragments， and then link the translation of each fragment， sentences that come together often appear as short sentences and as phrases. The sentences are not coherent and the translations cannot be properly adjusted according to the meaning of the sentence.
Secondly， the quality of translation differs greatly among different languages， different texts， and different genres. Although its translation system can translate many languages， the translation effects vary from language to language. Machine translation effect of informative text is much higher than any other emotional texts， operational.
texts， the translations of literature are not as artistic as the human translations.
2.2 The Merits and Demerits of Youdao Translation
Compared with Google translation， Youdao translation has become popular among young users， its success mainly depends on its powerful search engine （Youdao search）， and “web page extraction” technology， the translations mostly come from the web pages， so the translations can cover a lot of new vocabulary， and provide a variety of translation for the users to choose， which is the reason why the traditional translation tools are incomparable. Based on the results of web page extraction， the latest information can be included. However， there are still the following shortcomings：
Firstly， lack of dictionary source and corpus. Compared with the traditional translation tools， Youdao translation lacks unpopular and classic words.
Secondly， over-reliance on the Internet. Once out of the network， Youdao translation can not extract data from the page to achieve the desired translation results.
2.3 Take a business text as an example to compare the Google translation and Youdao translation Here takes one of the business texts as an instance， so it could be easier for us to know about the capabilities of current Google translation and Youdao translation.
Source text：A company’s total marketing communications program—called its promotion mix—consists of the specific blend of advertising， personal selling， sales promotion， and public relations tools that the company uses to pursue its advertising and marketing objectives.
Google Translation：公司的總营销传播计划—称其推广组合，包括广告，个人销售，促销和公关工具的具体组合，该公司用于追求其广告和营销目标。
Youdao Translation：一个公司的总营销传播项目—称为推广mix—包括广告、个人销售、促销和公关工具的具体组合，该公司利用这些工具来实现其广告和营销目标。
Human translation：公司的总体营销传播计划—即所谓的“促销组合”—包括该公司用以实现广告和营销目标的一系列工具的特定组合，这些工具包括广告宣传、人员推销、商品促销以及公共关系等等。
In the human translation， the translation of the original’s structure has been effectively adjusted. “that the company uses to pursue its advertising and marketing objectives” is regarded as an attributive， and this attributive is used to modify “tools”. While Google translation and Youdao translation failed to correctly distinguish which object the attributive was used to modify. In human translation， “tools” is translated as “一系列工具”， and put on the back of the attributive. “Specific blend” is translated into “特定组合”， which is put on the back of “a series of tools” to make the translation more fluent. However， the Google translation is similar to the human translation except that there is a defect in the grammatical analysis of the sentence. The Youdao translation is basically correct， but the special meaning of some specific words in certain contexts still cannot be properly analyzed， and it still have problems in the adjustment of the word orders.
3. Conclusion
In recent years， machine translation， represented by Google Translation and Youdao translation， has been actively explored users’ needs and tried to make its translations’ orders closer to the use of human language habits. Both have remarkable achievements. Google Translation’s Neural Machine Translation System （GNMT） enables machine translation to translate the texts that are closest to human translation. Youdao translation has positioned its main user groups as students and white-collar workers. Through its internet-based translation model， it can retrieve the latest words and translations that have not yet been included in the traditional dictionaries， which is favored by young groups. However， I believe that at present the two need to be improved in the following aspects： （1） Since Google translation uses a large-scale computing model and the basic idea of its system is to continue exploring linguistic patterns from a large number of samples by collecting samples， it is difficult to load a long text. The current Google translation can not translate the texts which beyond its ability. Such as English words， it can translate up to 5000 words at a time
（2） The current translation systems lack the awareness of judging contexts， so it is impossible to make clear judgments on the grammatical information such as tenses and subjunctive mood.In the translation of some special grammatical structures， the machine can not translate very well for lack of syntactic knowledge.
It is reported that the current machine translation performed well in the man-machine cooperation mode of translation work， and helped the translators save a lot of time to translate the repetitive academic words. In terms of the feasibility of machine translation， before the above problems remain unresolved， I am more inclined to the man-machine cooperation translation mode， that is， the machine translated the general ideas firstly， and then human translators proofread the translation again. It can help machine translation solve its problems in syntactic analysis and word order adjustment， and also provide useful guidance for machine translation development.
References：
[1]Yang Ming-xing，Yan Da.Study on translation model of diplomatic machine under modern information technology[J].Computer-assisted Foreign Language Education，2013（3）：33-41.
[2]Luo Jimei.Machine translation syntax error analysis[J].Journal of Tongji University（Social Science Section），2014，25（1）：111-118.
[3]Sun Jin.Machine translation based on the theory of text type[J].Chinese Science
【Key words】Machine Translation； Google Translation； Youdao Translation
1. Introduction
As human beings become more and more comprehensive about information， the quality of the human translation has also been greatly improved. However， with the explosive growth of the amount of information， the number of texts that need to be translated has also multiplied， and the supply of human translators is far from being able to meet the translation needs of massive translated texts. With the continuous efforts of the machine translation researchers over the years， the quality of machine translation is roughly the same as that of human beings， so the people’s confidence in using machine translation has been greatly improved. However， due to the fact that the current machine translation can not reasonably adjust the order of words and divide the sentence components， there is a big gap between machine translation and users’ expectations. Thus， the research on the possibility of machine translation is conducive to the development of machine translation to break through the bottleneck and achieve better development situation. It also has important a practical significance for the further exploration of translation field.
2. The Feasibility of Google Translation and Youdao Translation
The development of machine translation so far has made great progress， but there are also many deficiencies. Below will use Google translation and Youdao translation as examples to discuss the feasibility of the current machine translation. 2.1 The Advantages and Disadvantages of Google Translation
As the benchmark for the machine translation industry， Google Translation initially adopted the most popular mainstream method， Phrased-Based Machine Translation （PBMT）， which firstly cuts the sentence into phrases， and then matches the phrases from the prepared statistical database， find the most similar text， and then these phrases will be rearranged their orders in accordance with the target language’s syntax， then finally get the translation. The translations translated in this way often have the following problems：
Firstly， the translations are rigid， and even have syntactic confusion. Its translation method is to split the sentence into several fragments， and then link the translation of each fragment， sentences that come together often appear as short sentences and as phrases. The sentences are not coherent and the translations cannot be properly adjusted according to the meaning of the sentence.
Secondly， the quality of translation differs greatly among different languages， different texts， and different genres. Although its translation system can translate many languages， the translation effects vary from language to language. Machine translation effect of informative text is much higher than any other emotional texts， operational.
texts， the translations of literature are not as artistic as the human translations.
2.2 The Merits and Demerits of Youdao Translation
Compared with Google translation， Youdao translation has become popular among young users， its success mainly depends on its powerful search engine （Youdao search）， and “web page extraction” technology， the translations mostly come from the web pages， so the translations can cover a lot of new vocabulary， and provide a variety of translation for the users to choose， which is the reason why the traditional translation tools are incomparable. Based on the results of web page extraction， the latest information can be included. However， there are still the following shortcomings：
Firstly， lack of dictionary source and corpus. Compared with the traditional translation tools， Youdao translation lacks unpopular and classic words.
Secondly， over-reliance on the Internet. Once out of the network， Youdao translation can not extract data from the page to achieve the desired translation results.
2.3 Take a business text as an example to compare the Google translation and Youdao translation Here takes one of the business texts as an instance， so it could be easier for us to know about the capabilities of current Google translation and Youdao translation.
Source text：A company’s total marketing communications program—called its promotion mix—consists of the specific blend of advertising， personal selling， sales promotion， and public relations tools that the company uses to pursue its advertising and marketing objectives.
Google Translation：公司的總营销传播计划—称其推广组合，包括广告，个人销售，促销和公关工具的具体组合，该公司用于追求其广告和营销目标。
Youdao Translation：一个公司的总营销传播项目—称为推广mix—包括广告、个人销售、促销和公关工具的具体组合，该公司利用这些工具来实现其广告和营销目标。
Human translation：公司的总体营销传播计划—即所谓的“促销组合”—包括该公司用以实现广告和营销目标的一系列工具的特定组合，这些工具包括广告宣传、人员推销、商品促销以及公共关系等等。
In the human translation， the translation of the original’s structure has been effectively adjusted. “that the company uses to pursue its advertising and marketing objectives” is regarded as an attributive， and this attributive is used to modify “tools”. While Google translation and Youdao translation failed to correctly distinguish which object the attributive was used to modify. In human translation， “tools” is translated as “一系列工具”， and put on the back of the attributive. “Specific blend” is translated into “特定组合”， which is put on the back of “a series of tools” to make the translation more fluent. However， the Google translation is similar to the human translation except that there is a defect in the grammatical analysis of the sentence. The Youdao translation is basically correct， but the special meaning of some specific words in certain contexts still cannot be properly analyzed， and it still have problems in the adjustment of the word orders.
3. Conclusion
In recent years， machine translation， represented by Google Translation and Youdao translation， has been actively explored users’ needs and tried to make its translations’ orders closer to the use of human language habits. Both have remarkable achievements. Google Translation’s Neural Machine Translation System （GNMT） enables machine translation to translate the texts that are closest to human translation. Youdao translation has positioned its main user groups as students and white-collar workers. Through its internet-based translation model， it can retrieve the latest words and translations that have not yet been included in the traditional dictionaries， which is favored by young groups. However， I believe that at present the two need to be improved in the following aspects： （1） Since Google translation uses a large-scale computing model and the basic idea of its system is to continue exploring linguistic patterns from a large number of samples by collecting samples， it is difficult to load a long text. The current Google translation can not translate the texts which beyond its ability. Such as English words， it can translate up to 5000 words at a time
（2） The current translation systems lack the awareness of judging contexts， so it is impossible to make clear judgments on the grammatical information such as tenses and subjunctive mood.In the translation of some special grammatical structures， the machine can not translate very well for lack of syntactic knowledge.
It is reported that the current machine translation performed well in the man-machine cooperation mode of translation work， and helped the translators save a lot of time to translate the repetitive academic words. In terms of the feasibility of machine translation， before the above problems remain unresolved， I am more inclined to the man-machine cooperation translation mode， that is， the machine translated the general ideas firstly， and then human translators proofread the translation again. It can help machine translation solve its problems in syntactic analysis and word order adjustment， and also provide useful guidance for machine translation development.
References：
[1]Yang Ming-xing，Yan Da.Study on translation model of diplomatic machine under modern information technology[J].Computer-assisted Foreign Language Education，2013（3）：33-41.
[2]Luo Jimei.Machine translation syntax error analysis[J].Journal of Tongji University（Social Science Section），2014，25（1）：111-118.
[3]Sun Jin.Machine translation based on the theory of text type[J].Chinese Science